Re: The history of stdint.h ?

[Garance A Drosihn <drosih@xxxxxxx>]

At 2:22 PM -0700 6/19/04, walt wrote:
> I'm trying to debug the grub port, which fails with error messages
> about previous definitions of things previously defined in stdint.h.
>
> I see that stdint.h is defined in DFlyBSD and FBSD5, but not FBSD4.
>
> So I assume that stdint.h was MFC'd from FBSD5 to DFly -- is this
> right or wrong?
>
> I also have read that stdint.h is part of gnu libc -- is this
> right or wrong?  If so, why was it adopted as part of FBSD5 but
> not FBSD4?

I think this all boils down to "4.x-stable is old".  Not only is
it old, but it has been "stable" for a long time.  "Stable" means
that ABI's and API's should not see dramatic changes.  Sorting
out variables into stdint.h is just too big of a change for a
stable branch -- particularly since we HOPE to make 5.x the
stable branch very soon now.

While dragonfly is based on 4.x-stable, it does not have the same
requirement to avoid major changes.  It is not a stable branch
being used by many people in a production environment.  Matt &
company can feel free to change anything and everything in
dragonfly without disrupting some end-users who are depending on
a very stable (aka "static, unchanging") environment.

I have no idea what you mean by "stdint.h being part of gnu's
libc".  libc is an object library.  It is not an include file.
The FreeBSD and Dragonfly versions of stdint.h are certainly
not GNU-licensed.  stdint.h is part of the latest POSIX, and
part of SingleUnixSpec-version 3.  Any modern unix should want
to include it, once they take the time to make sure that the
version they implement will conforms to the standards.

--
Garance Alistair Drosehn            =   gad@xxxxxxxxxxxxxxxxxxxx
Senior Systems Programmer           or  gad@xxxxxxxxxxx
Rensselaer Polytechnic Institute    or  drosih@xxxxxxx